update inference

code/inference.py

@@ -2,7 +2,7 @@ import requests
 from PIL import Image
 from io import BytesIO
 import torch
-from transformers import AutoTokenizer, BitsAndBytesConfig
+from transformers import AutoTokenizer
 
 from llava.model import LlavaLlamaForCausalLM
 from llava.utils import disable_torch_init

code/requirements.txt

@@ -1,2 +1 @@
-# llava @ git+https://github.com/haotian-liu/LLaVA@main
 llava @ git+https://github.com/haotian-liu/[email protected]